The document discusses distributed computing systems, highlighting their inherent nature and requirements, such as information sharing, resource sharing, and improved performance metrics over centralized systems. It explains various aspects like reliability, extensibility, and flexibility, comparing network and distributed operating systems based on features like transparency and fault tolerance. The content also addresses the importance of scalability and performance optimization principles in the design of distributed systems.

1. Presented By
Dr. A. ASHOK KUMAR,
Assistant Professor,
Department Of Computer Science,
Alagappa Government Arts College,
Karaikudi – 630003.
ashokamjuno@rediffmail.com

2. 1. Inherently Distributed Applications
 several applications are inherently distributed in nature and
require a distributed computing system
 an employee database of a nationwide organization, the data to
a particular employee are generated at the employee's branch
office
• to the global need to view the entire database
• a local need for frequent and immediate access to locally
generated data at each branch office
 some processing power be available at the many distributed
locations for collecting, preprocessing, and accessing data,
resulting in the need for distributed computing systems
 other examples airline reservation system, a computerized
banking system

3. 2. Information Sharing among Distributed Users
 person-to-person communication facility by sharing
information
 information generated by one of the users can be easily and
efficiently shared by the users working at other nodes of the
system
 For example, a project can be performed by two or more users
who are geographically far off
 The use of distributed computing systems by a group of users
to work cooperatively is known as computer-supported
cooperative working (CSCW), or groupware
 Groupware applications depend heavily on the sharing of
data objects between programs running on different nodes
of a distributed computing system
 Groupware is an emerging technology for software
developers

4. 3. Resource Sharing
 Sharing of software resources such as software
libraries and databases
 hardware resources such as printers, hard disks,
and plotters
 a distributed computing system based on the
workstation-server model
4. Better Price-Performance Ratio
 rapidly increasing power and reduction in the
price of microprocessors, combined with the
increasing speed of communication networks
 distributed computing systems potentially have a
much better price-performance ratio than a single
large centralized system

5.  The processor-pool model can be effectively used
by a large number of users from inexpensive
terminals
 Facilitate resource sharing among multiple
computers
 A single unit of expensive peripheral devices such
as color laser printers, high-speed storage devices,
and plotters can be shared

6. 5. Shorter Response times and Higher Throughput
 Due to multiplicity of processors, distributed
computing systems are expected to have better
performance than single-processor centralized
systems
 The two most commonly used performance metrics
are
• response time and
• throughput of user processes
 computation can be partitioned into a number of
subcomputations that can run concurrently
 DCS with very fast communication networks are
increasingly being used as parallel computers to solve
single complex problems rapidly

7. 6. Higher Reliability
 Reliability refers to the degree of tolerance
against errors and component failures in a
system
 A reliable system prevents loss of information
even in the event of component failures
 The multiplicity of storage devices and
processors increase reliability
• if one of the processors fails, the computation
can be successfully completed at the other
processor
• if one of the storage devices fails, the
information can still be used from the other
storage device

8.  An important aspect of reliability is availability, which
refers to the fraction of time for which a system is
available for use
• if the processor of a centralized system fails the entire
system breaks down and no useful work can be
performed
• distributed computing system, a few parts of the system
can be down without interrupting the jobs of the users
 workstation-server model fails, only the user of that
workstation is affected
 the processor-pool model, if some of the processors in
the pool are down at any moment, the system can
continue to function normally

9. 7. Extensibility and Incremental Growth
 it is possible to gradually extend the power and
functionality of a distributed computing system by
simply adding additional resources
 Additional processors can be easily added to the
system to handle the increased workload
 existing and proposed applications it is practically
impossible to predict future demands of the
system
 distributed computing systems that have the
property of extensibility and incremental growth
are called open distributed systems

10. 8. Better Flexibility in Meeting Users' Needs
 Different types of computers are usually more suitable for
performing different types of computations
 computers with ordinary power are suitable for ordinary data
processing jobs
 whereas high-performance computers are more suitable for
complex mathematical computations
 In a centralized system, the users have to perform all types of
computations on the only available computer
 In a distributed computing system from the different types of
computers, the most appropriate one selected for processing a
user's job depending on the nature of the job
 interactive jobs can be processed at a user's. own workstation
 the processors in the pool may be used to process
noninteractive

11.  An operating system as a program that controls
• the resources of a computer system and
• provides its users with an interface
 the two primary tasks of an operating system are
• To present users with a virtual machine that is easier to
program than the underlying hardware
• To manage the various resources of the system
 The OS commonly used for distributed computing systems can
be broadly classified into two types
• Network operating systems and
• distributed operating systems
 The three most important features commonly used to
differentiate between these two types

12. 1. System image
Network Operating System Distributed operating
system
1 the users view the distributed
computing system as a collection of
distinct machines connected by a
communication subsystem
hides the existence of multiple
computers and provides a single-
system image to its users
it makes a collection of networked
machines act as a virtual
uniprocessor
2 a user is required to know the
location of a resource to access it
need not keep track of the
locations of various resources for
accessing them
3 different sets of system calls have to
be used for accessing local and
remote resources
the same set of system calls is used
for accessing both local and
remote resources

13. 2. Autonomy
Network Operating System Distributed operating system
1 built on a set of existing centralized
operating systems handles the
interfacing and coordination of remote
operations and communications
between these operating systems
single system wide operating system and
each computer of the distributed
computing system runs a part of this global
operating system
2 each computer of the distributed
computing system has its own local
operating system
when two processes of different
computers communicate with each
other, they must use a mutually agreed
on communication protocol
they work in close cooperation with each
other for the efficient and effective
utilization of the various resources of the
system
3 Each computer functions
independently of other computers
processes and several resources are
managed globally (some resources are
managed locally)
a single set of globally valid system calls
available on all computers of the
distributed computing system

14.  The set of system calls that an operating system
implemented by a set of programs called the kernel
of the operating system
 The kernel manages and controls the hardware of
the computer system
 system calls globally valid, with a distributed
operating system identical kernels are run on all the
computers
 The kernels of different computers often cooperate
with each other in making global decisions
 The degree of autonomy in distributed computing
system that uses a network operating system is high
as compared to distributed operating system

15. 3. Fault tolerance capability
 A network operating system provides little or no
fault tolerance capability
 distributed operating system, most of the users
are normally unaffected by the failed machines
and can continue to perform their work normally
 the fault tolerance capability of a distributed
operating system is usually very high as
compared to that of a network operating system
 A distributed computing system that uses a
network operating system is usually referred to as
a network system
 one that uses a distributed operating system is
usually referred to as a true distributed system

16. 1. Transparency
To make the existence of multiple computers invisible
(transparent)
Provide a single system image to its users
Collection of distinct machines connected by a
communication subsystem appears to its users as a
virtual uniprocessor
Achieving complete transparency is a difficult task
Several different aspects of transparency be supported
by the distributed operating system
8 forms of transparency identified by the International
Standards Organization’s Reference Model for Open
Distributed Processing
Access, Location, replication, failure, migration,
concurrency, performance and scaling transparency

17. Access transparency
It means that users should not need or be able to
recognize whether a resource is remote or local
The DOS allow users to access remote resources in
the same way as local resources
The user interface takes the form of a set of system
calls
It should not distinguish between local and remote
resources
It should be the responsibility of DOS to locate the
resources and to arrange for servicing user requests
in a user-transparent manner

18. Location transparency
Two main aspects of location transparency are
Name transparency
 The name of the resource should not reveal any hint as to
the physical location of the resource
 The name of the resource is independent of the physical
connectivity or topology of the system or the current
location of the resource
 The resource name must be unique systemwide
User mobility
 Fact that no matter which machine a user is logged onto
 User should be access a resource with the same name
 i.e the user should not be required to use different names
to access the same resource from two different nodes
 User can freely log on to any machine in the system and
access any resource without making extra effort.

19.  Replication transparency
For better performance and reliability all DOS have
provision to create replicas of files and other resources
on different nodes
The existence of multiple copies of replicated resource
and replication activity to be transparent to users
Two important issues are
Naming of replicas
 Responsibility of the system to name the various copies of
resources
 To map a user-supplied name of the resource to an appropriate
replica of the resource
Replication control
 How many copies of the resource to be created
 Where should each copy to be placed
 When should a copy be created /deleted automatically in a
user-transparent manner

20. Failure transparency
Deals with masking from the users partial failure in
the system
Such as a communication link failure, a machine
failure, or a storage device crash
A DOS with this property will continue to function,
in a degraded form, in the face of partial failure
Example of file server failure in case of file service
Complete transparency is not achievable
Failure of communication network of a disrupts the
work of its users and is noticeable by the users

21. Migration transparency
For better performance, reliability, and security
reasons , an object is capable of being moved
It is often migrated from one node to another in a
distributed system
This ensure that the movement of the object is
handled automatically by the system in a user-
transparent manner
Three important issues are
 Migration decisions – which object is to be moved from
where to where should be made automatically by the
system
 Migration of an object from one node to another should
not require any change in its name
 When migration object is a process – the IPC machanism
ensure that a message sent to the migrating process
reaches it with out the need of resend it

22. Concurrent transparency
Multiple users who are spatially separated use the
system concurrently
It is economical to share the system resources
among the concurrently executing user processes
Number of available resources in a computing
system is restricted
One user process influence the action of another
process as it competes for resources
Four properties are
 An event ordering property ensures that all access
request to various system resources are properly ordered
 A mutual exclusion property
 A no-starvation property
 A no-deadlock property

23. Performance transparency
To allow system to be automatically reconfigured to
improve performance , as load vary dynamically in
the system
One processor is overloaded with jobs while another
processor is idle should not be allowed to occur
The processing capability should be uniformly
distributed among the currently available jobs
This uses the intelligent resource allocation and
process migration facilities
Scaling transparency
To allow the system to expand in scale without
disrupting the activities of the users
Use the scalable algorithms for designing the DOS
components

24. 2. Reliability
 Distributed systems are more reliable than
centralized systems due to the existence of multiple
instances of resources
 Multiple resources alone cannot increase the
system’s reliability
 A fault is a mechanical or algorithmic defect that may
generate an error
 A fault in a system causes system failure
 System failures are two types:
Fail-stop failure – system stops functioning after changing to a
state in which its failure can be detected
Byzantine failure – the system continues to function but
produces wrong results
 Byzantine failures are more difficult to deal with than fail-stop
failures
 Fault handling mechanisms designed properly to
avoid faults, to tolerate faults, and to detect and
recover from faults

25.  Commonly used methods used for faults is
 Fault avoidance
Deals with designing the components of the system in
such a way that the occurrence of faults is minimized
High reliability components are often employed for
improving the system reliability
Software components test the hardware components to
make these components highly reliable
 Fault Tolerance
Ability of a system to continue functioning in the event of
partial system failure
Performance degraded but the system functions
properly
Concepts to improve the fault tolerance ability is:
Redundancy techniques
 To avoid single point of failure by replicating critical hardware
and software components
 Additional system overhead is needed to maintain two or more
copies of a replicated resource
 Keep all copies of a resource consistent

26.  How much replica is enough?
 K-fault tolerant if it can continue to function even in the event
of the failure of k components
 If the system is to be designed to tolerate k fail-stop failures,k+1
replicas are needed
 If the system is tolerate k byzantine failures, a minimum of 2k+1
replicas are needed
 Because a voting mechanism to be used the majority k+1 of the
replicas when k replicas behave abnormally
Distributed control
 Distributed control mechanism to avoid single points of
failure
 Highly available distributed file system have multiple and
independent file servers controlling multiple and
independent storage devices
 It is also used for name servers, scheduling algorithms,
and other executive control functions

27. Fault Detection and Recovery
Commonly used methods are
Atomic transactions
 Computation consisting of operations take place
indivisibly in the presence of failures and concurrent
computations
 Either all of the operations performed successfully or
none of their effects prevails
 Other processes executing concurrently cannot modify or
observe intermediate state of the computation
 If a process halts unexpectedly due to a hardware faults
or a software fault before a transaction is completed
 The system subsequently restores any data objects that
were undergoing modification to their original states
Stateless servers
 Two service paradigms: stateful or stateless

28.  Stateful approach depend on the history of the serviced
requests
 Stateless approach does not depend on it
 Stateless server makes crash recovery very easy because
no client information is maintained by server
 Stateful paradigm requires complex crash recovery
procedures
 Both server and client need to reliably detect crashes
Acknowledgement and timeout-based
retransmissions of messages
 Retransmission based on acknowledgement and timeouts
 Receiver return acknowledgement
 Within the fixed timeout period message is re transmitted
 It results duplicate messages
 Need mechanism for handling duplicate messages using
sequence numbers

29. 3. Flexibility
Most important feature for open distributed
systems
Flexibility due to the following reasons
Ease of modification
 System designers often need to be replaced/modified
either because some bug is detected
 The dsign is no longer suitable for the changed system
environment or new-user requirements
 Incorporate changes in the system in a user-transparent
manner or with minimum interruption caused to the user
Ease of enhancement
 New functionalities have to be added from time to time to
make it more powerful and easy to use
 User group has the flexibility to add their own services

30.  Flexibility of a distributed OS is the model used for
designing kernel
 The kernel of an OS is its central controlling part that
provides the system facilities
 It operates in a separate address space that is
inaccessible to user processes
 It is the only part of the OS that a user cannot replace or
modify
 The two commonly used models for kernel
Monolithic kernel
 Most operating system services such as process management, memory
management, device management, file management, name
management, and interprocess communication are provided by the
kernel
 Unix is monolithic type
Micro kernel
 The kernel is very small nucleus of software that provides only the
minimal facilities
 Only service provided in this model is IPC, low-level device
management, low-level process management, and some memory
management
 Other services implemented as user level processes

31. 4. Performance
 To achieve performance various components of
the distributed system be designed properly
 Some design principles are
Batch if possible
 Transfer of data across the network in large chunks rather than
individual pages
 Piggybacking of acknowledgement of previous message with the
next message.
Cache whenever possible
 Caching of data at clients sites frequently improves overall system
performance
 It makes the data available
 Saves large amount of computing time and network bandwidth
Minimize copying of data
 To avoid copying of data, to make optimal use of memory
management
 It helps eliminating data movement between the kernel, block I/O
devices, clients and servers

32. Minimize network traffic
 Performance also improved by reducing internode
communication costs
 Access to remote resources require communication
through intermediate nodes
Migrating a process to closer to the resources it is using will
reduce network traffic
Cluster two or more processes that frequently communicate
with each other on the same node of the system
Take advantage of fine-grain parallelism for
multiprocessing
 Threads used for structuring server process
 Concurrency control of simultaneous access by multiple
processes to a shared resources

33. 5. Scalability
 Refers to the capability of a system to adopt to increased
service load
 Distributed system will grow
 DOS should be designed to easily cope with the growth
of nodes and users in the system
 It should not cause disruption of service or significant
loss of performance
 Some principles for designing scalable distributed
systems
Avoid centralized entities
 Use of centralized entities such as a single file server or a single
database for the entire system makes nonscalable
a) The failure of the centralized entity brings the entire system down
b) Performance becomes a system bottleneck when contention for it
increases
c) Connection between centralized entity with other nodes gets saturated
d) Several interconnected LAN is inefficient to a particular type of request
serviced at a server node

34.  Avoid centralized algorithms
 Centralized algorithm operates by collecting information from all nodes, processing this
information on a single node and then distributing the results to other nodes
 Decentralized algorithms used , need not collect data
 It uses locallaly available information
 Performa most operations on client workstations
 An operation should be performed on the clients own workstation rather than on a
server machine
 Server is a common resource for several clients and hence server cycles are more
precious than the cycles of client workstations
6. Heterogeneity
 Designing heterogeneous distributed systems is more difficult than
designing homogeneous systems
 Heterogeneous system needs some form of translation is necessary
for interaction between two incompatible nodes.
 Data translation performed at the sender’s node or at the receivers
nodes
 This translation process can be greatly reduced by using
intermediate standard data format
 Each node requires a translation software for converting from its
own format to standard format, and from standard format to its own
format.

35. 7. Security
 Security is difficult because of the lack of a single point of
control
 Use of insecure networks for data communication
 In a centralized system all users are authenticated by the
system at login time
 In a distributed system server should know who is the client
request for service
 Client identification field in the message cannot be trusted
 Because an intruder may change or act as a client during
transmission
 Distributed system has the following requirements
Sender should be know that the message was received by the
intended receiver
Receiver should know the message was sent by the genuine
sender
Sender and receiver guaranteed that message were not changed
during transfer
8. Emulation of existing Operating system
 Moving to the new DOS will allow both types of software to
be run side by side

36. Introduction To Distributing Computing
Environment (DCE)
 DCE was defined by the Open Software
Foundation(OSF)
 It is a consortium of computer manufacturers
 It is not an OS, nor is it an application
 It is an integrated set of services and tools on
the top of existing OS
 Serve as a platform for building and running
distributed applications
 A primary goal of DCE is vendor independence
 It runs on different kinds of computers, OS, and
networks
 DCE is a middleware software layered between
the DCE applications layer and the operating
system and networking layer

37.  Each machine has its own local OS
 The DCE software layer on top of the OS and
networking layer
DCE applications
DCE software
Operating systems and
networking

38. DCE Components
 Thread package
Provides a simple programming model for building
concurrent applications
It includes operations to create and control multiple threads
of execution in a single process
Synchronize access to global data within the application
 Remote Procedure Call (RPC)
It provide tools to build client-server applications
DCE RPC is basis for all communication in DCE
It is ease to use
It is network- and protocol-independent
Provides secure communication between client and server
Hides differences in data requirements by converting data to
appropriate format

39.  Distributed Time Service (DTS)
It synchronize the clocks of all the computers in the system
Also permits the use of time values from external time
sources
It synchronize the clocks of computers in the system with
external time
Used to synchronize the clocks of the computers of one
distributed environment with another.
 Name services
Includes Cell Directory service (CDS), Global Directory Service
(GDS), and the Global directory Agent (GDA)
These allows servers, files, devices to be uniquely named and
accessed in a location-transparent manner

40.  Security service
 Provides tools for authentication and authorization to protect
system resources
 Distributed File Service (DFS)
 It provides systemwide file system
 It has characteristics such as location transparency, high
performance, and high availability
 Unique feature of DCE DFS is also provide file services to clients of
other file systems
 The DCE components are tightly integrated
 Interdependencies of DCE components are recursive
Component Name Other Components Used
by It
Threads None
RPC Threads, name, security
DTS Threads, RPC, name,
security
Name Threads, RPC, DTS,
Security
Security Threads, RPC, DTS, name

41. DCE cells
 A cell is a group of users, machines, or other
resources
 The minimum cell configuration requires a cell
directory server, a security server, a distributed
time server and one or more client machines
 Each DCE client machine has client processes for
security service, cell directory service, distributed
time service, RPC facility, and threads facility
 Setting up a DCE system is to decide cell
boundaries
 Four factors to be considered for decision making
Purpose
 Users working on the same goal should be put in the same cell
 They need easy access to a common set of system resources

42. Administration
 Each system needs an administrator to register new users in
the system
 And decide the access rights to the system resources
 All the machines and their users manageable by a
administrator put in a single cell
Security
 Machines of those users who have greater trust in each other
should be put in the same cell
 Cell boundaries act like a firewalls that accessing resources
belongs to another cell
Overhead
 Name resolution, user authentication incur overhead when
they performed between cells than when they are performed
within the same cell
 Machines of users who frequently interact with each other and
the resources accessed frequently placed in same cell